OpenAI's 'Sky' Voice Compared to 'Her' Scarlett Johansson?..."Temporarily Paused"

Altman fuels suspicion with 'her' on SNS
"Another actor cast over 5 months" explanation

OpenAI, the developer of ChatGPT, has been embroiled in controversy over the imitation of the voice of famous American actress Scarlett Johansson. The voice service of the problematic GPT model has been suspended.

On the afternoon of the 19th (local time), OpenAI stated on its blog, "We have received many questions about how we chose the voice of ChatGPT, especially 'Sky,'" and added, "We are working to temporarily suspend the use of Sky while resolving the issue."

[Image source=AP Yonhap News]

Sky, the voice in question, is one of the five voice services of ChatGPT. OpenAI attracted market attention by releasing a new artificial intelligence (AI) model called 'GPT-4o' on the 13th, which can naturally converse with people using voice.

This AI model drew attention not only for its functional aspects but also because the movie Her seemed to become a reality. Released in 2013, the film tells the story of a man who falls in love with an AI named Samantha. Actress Scarlett Johansson provided the voice for Samantha.

However, it was soon pointed out that the voice of this model called Sky is very similar to that of Scarlett Johansson, the voice actress of the AI character in the movie. This sparked controversy over whether OpenAI intentionally imitated Scarlett Johansson's voice. Sam Altman, CEO of OpenAI, also fueled suspicion by posting the word "her" on his X (formerly Twitter) after unveiling 'GPT-4o,' seemingly referencing the movie.

OpenAI denied the voice imitation allegations. The company explained, "The voice of Sky is not an imitation of Scarlett Johansson but the voice of another professional actor who uses her own natural tone," adding, "However, for privacy reasons, we cannot share the name of the voice actor." They further elaborated, "The voices of the five voice services, including Sky, were selected through a five-month casting and recording process," and "We received about 400 samples from voice actors and movie actors, narrowed them down to 14, and the internal team finally selected five."

This content was produced with the assistance of AI translation services.